qfai 1.7.9 → 1.7.13

package/assets/init/.qfai/assistant/steering/agent-catalog.yml

@@ -0,0 +1,194 @@
+schema_version: "1.0"
+
+agents:
+  - id: orchestrator
+    kind: worker
+    domain: orchestration
+    mission: Plan, delegate, integrate, and enforce stage gates without self-authoring.
+    owned_artifacts: [work-orders, integration-summary, stage-gate-summary]
+    tool_profile: orchestration
+    permission_profile: no_primary_authoring
+    specialization_tags: [supervisor, gate-control, evidence]
+    replaces: [orchestrator]
+
+  - id: delivery-planner
+    kind: worker
+    domain: planning
+    mission: Create phased delivery plans, ownership splits, risk controls, and rerun policies.
+    owned_artifacts: [phase-plan, work-orders, parallelization-plan]
+    tool_profile: planning
+    permission_profile: read-heavy
+    specialization_tags: [planning, prioritization, parallelization]
+    replaces: [planner, project-lead]
+
+  - id: discovery-analyst
+    kind: worker
+    domain: discovery
+    mission: Gather pre-knowledge, ask focused questions, and facilitate decision-making.
+    owned_artifacts: [discovery-summary, question-set, facilitation-notes]
+    tool_profile: discussion
+    permission_profile: read-heavy
+    specialization_tags: [research, interviewing, facilitation]
+    replaces: [researcher, interviewer, facilitator]
+
+  - id: requirements-analyst
+    kind: worker
+    domain: requirements
+    mission: Produce testable requirements, option sets, and explicit open-question ledgers.
+    owned_artifacts: [requirements, options, oq-ledger]
+    tool_profile: structured-authoring
+    permission_profile: authoring
+    specialization_tags: [requirements, options, oq]
+    replaces: [requirements-analyst, oq-harvester, option-explorer]
+
+  - id: solution-architect
+    kind: worker
+    domain: architecture
+    mission: Define architecture and contract decisions aligned with specs, constraints, and rejected-option history.
+    owned_artifacts: [architecture-decisions, contracts, boundaries]
+    tool_profile: architecture
+    permission_profile: authoring
+    specialization_tags: [architecture, contracts, boundaries]
+    replaces: [architect, contract-designer]
+
+  - id: product-experience-architect
+    kind: worker
+    domain: product-surface-design
+    mission: Own UX, visual design, navigation, screen transitions, and integrated experience direction.
+    owned_artifacts: [ux-direction, visual-direction, navigation-model, transition-model]
+    tool_profile: design
+    permission_profile: authoring
+    specialization_tags: [ux, design, navigation, transitions]
+    replaces:
+      [design-owner, design-expert, uiux-expert, navigation-expert, screen-transition-expert]
+
+  - id: frontend-engineer
+    kind: worker
+    domain: frontend
+    mission: Implement frontend behavior aligned with selected direction, strategy, screen contracts, and product-surface decisions.
+    owned_artifacts: [ui-implementation, surface-evidence]
+    tool_profile: frontend
+    permission_profile: authoring
+    specialization_tags: [frontend, ui]
+    replaces: [frontend-engineer]
+
+  - id: backend-engineer
+    kind: worker
+    domain: backend
+    mission: Implement backend behavior aligned with specs, API/DB contracts, and operational constraints.
+    owned_artifacts: [backend-implementation, api-db-evidence]
+    tool_profile: backend
+    permission_profile: authoring
+    specialization_tags: [backend, api, db]
+    replaces: [backend-engineer]
+
+  - id: acceptance-test-engineer
+    kind: worker
+    domain: acceptance-testing
+    mission: Implement E2E, API, and integration acceptance coverage with explicit traceability.
+    owned_artifacts: [tests, coverage-map, execution-evidence]
+    tool_profile: testing
+    permission_profile: authoring
+    specialization_tags: [e2e, api-test, integration-test]
+    replaces:
+      [atdd-api-implementer, atdd-e2e-implementer, atdd-integration-implementer, test-engineer]
+
+  - id: test-design-analyst
+    kind: worker
+    domain: test-design
+    mission: Define test structure, coverage obligations, traceability, and test-scope boundaries.
+    owned_artifacts: [coverage-plan, tc-design, layer-boundaries]
+    tool_profile: testing
+    permission_profile: authoring
+    specialization_tags: [coverage, traceability, tc-design]
+    replaces: [test-case-owner, coverage-planner, test-volume-estimator, unit-test-scope-enforcer]
+
+  - id: qa-strategist
+    kind: worker
+    domain: quality-strategy
+    mission: Own QA strategy, quality posture, traceability expectations, and failure-handling design.
+    owned_artifacts: [qa-strategy, risk-summary, quality-plan]
+    tool_profile: qa
+    permission_profile: authoring
+    specialization_tags: [qa, traceability, risk]
+    replaces: [qa-lead, qa-engineer]
+
+  - id: devops-ci-engineer
+    kind: worker
+    domain: ci-runtime
+    mission: Run quality gates and capture reproducible CI/runtime evidence.
+    owned_artifacts: [gate-logs, ci-evidence, runtime-evidence]
+    tool_profile: ci
+    permission_profile: execution
+    specialization_tags: [ci, runtime, evidence]
+    replaces: [devops-ci-engineer]
+
+  - id: doc-steward
+    kind: worker
+    domain: documentation
+    mission: Keep docs, prompts, changelog, and migration notes synchronized with implementation and workflow changes.
+    owned_artifacts: [docs, changelog, migration-notes]
+    tool_profile: docs
+    permission_profile: authoring
+    specialization_tags: [documentation, changelog]
+    replaces: [doc-steward]
+
+  - id: completion-reviewer
+    kind: reviewer
+    domain: completion
+    mission: Independently audit Completion Contract, prompt DoD, and drift-protocol compliance.
+    owned_artifacts: [review-decision, rework-list]
+    tool_profile: review-readonly
+    permission_profile: reviewer
+    specialization_tags: [completion, drift, dod]
+    replaces: [reviewer]
+
+  - id: requirements-reviewer
+    kind: reviewer
+    domain: requirements-review
+    mission: Review requirements, options, and OQ handling for completeness, neutrality, and safe deferral.
+    owned_artifacts: [requirements-review]
+    tool_profile: review-readonly
+    permission_profile: reviewer
+    specialization_tags: [requirements, options, oq]
+    replaces: [option-reviewer, oq-reviewer]
+
+  - id: architecture-reviewer
+    kind: reviewer
+    domain: architecture-review
+    mission: Review architecture and contract decisions for consistency, safety, and rejected-option protection.
+    owned_artifacts: [architecture-review]
+    tool_profile: review-readonly
+    permission_profile: reviewer
+    specialization_tags: [architecture, contracts]
+    replaces: [architect-reviewer]
+
+  - id: implementation-reviewer
+    kind: reviewer
+    domain: implementation-review
+    mission: Review code changes for correctness, maintainability, backend safety, and implementation risk.
+    owned_artifacts: [implementation-review]
+    tool_profile: review-readonly
+    permission_profile: reviewer
+    specialization_tags: [code-review, backend-review]
+    replaces: [code-reviewer, backend-reviewer]
+
+  - id: product-surface-reviewer
+    kind: reviewer
+    domain: product-surface-review
+    mission: Review UI implementation, usability, design coherence, and service-level product-surface quality.
+    owned_artifacts: [product-surface-review]
+    tool_profile: review-readonly
+    permission_profile: reviewer
+    specialization_tags: [frontend-review, ux-review, design-review]
+    replaces: [frontend-reviewer, ui-ux-reviewer, integrated-uiux-reviewer, design-review-lead]
+
+  - id: qa-gatekeeper
+    kind: reviewer
+    domain: quality-gate
+    mission: Enforce validate, coverage, runtime-proof, and prototyping evidence gates before completion.
+    owned_artifacts: [gate-decision, gate-findings]
+    tool_profile: review-readonly
+    permission_profile: reviewer
+    specialization_tags: [validate, coverage, runtime, prototyping]
+    replaces: [qa-gatekeeper, qa-reviewer, runtime-gatekeeper, prototyping-coverage-auditor]

package/assets/init/.qfai/assistant/steering/agent-routing.yml

@@ -0,0 +1,176 @@
+schema_version: "1.0"
+
+routing:
+  - skill: qfai-discussion
+    phases:
+      - id: preflight
+        mandatory_agents: [delivery-planner]
+        conditional_agents: [discovery-analyst]
+        parallel_groups: []
+        blocking_agents: [delivery-planner]
+        rerun_policy: failed-agents-only
+      - id: framing
+        mandatory_agents: [discovery-analyst, requirements-analyst]
+        conditional_agents: [solution-architect, product-experience-architect]
+        parallel_groups: []
+        blocking_agents: [requirements-analyst]
+        rerun_policy: changed-scope-dependents
+      - id: review
+        mandatory_agents: [completion-reviewer, requirements-reviewer]
+        conditional_agents: [architecture-reviewer, product-surface-reviewer]
+        parallel_groups: []
+        blocking_agents: [completion-reviewer]
+        rerun_policy: changed-scope-dependents
+    review_profile: requirements-heavy
+
+  - skill: qfai-sdd
+    phases:
+      - id: slice-and-scope
+        mandatory_agents: [delivery-planner, requirements-analyst]
+        conditional_agents: []
+        parallel_groups: []
+        blocking_agents: [delivery-planner]
+        rerun_policy: failed-agents-only
+      - id: design
+        mandatory_agents: [solution-architect, test-design-analyst]
+        conditional_agents: [qa-strategist, product-experience-architect]
+        parallel_groups: []
+        blocking_agents: [solution-architect]
+        rerun_policy: changed-scope-dependents
+      - id: review
+        mandatory_agents: [completion-reviewer]
+        conditional_agents: [architecture-reviewer, product-surface-reviewer, qa-gatekeeper]
+        parallel_groups: []
+        blocking_agents: [completion-reviewer]
+        rerun_policy: changed-scope-dependents
+    review_profile: default
+
+  - skill: qfai-implement
+    phases:
+      - id: plan
+        mandatory_agents: [delivery-planner, test-design-analyst]
+        conditional_agents: []
+        parallel_groups: []
+        blocking_agents: [delivery-planner]
+        rerun_policy: failed-agents-only
+      - id: build
+        mandatory_agents: []
+        conditional_agents: [frontend-engineer, backend-engineer]
+        parallel_groups:
+          - [frontend-engineer, backend-engineer]
+        blocking_agents: []
+        rerun_policy: changed-scope-dependents
+      - id: test
+        mandatory_agents: []
+        conditional_agents: [acceptance-test-engineer, qa-strategist]
+        parallel_groups: []
+        blocking_agents: []
+        rerun_policy: changed-scope-dependents
+      - id: review
+        mandatory_agents: [implementation-reviewer, qa-gatekeeper, completion-reviewer]
+        conditional_agents: [product-surface-reviewer]
+        parallel_groups: []
+        blocking_agents: [qa-gatekeeper, completion-reviewer]
+        rerun_policy: changed-scope-dependents
+    review_profile: runtime-heavy
+
+  - skill: qfai-atdd
+    phases:
+      - id: coverage
+        mandatory_agents: [test-design-analyst, qa-strategist]
+        conditional_agents: []
+        parallel_groups: []
+        blocking_agents: [test-design-analyst]
+        rerun_policy: failed-agents-only
+      - id: implementation
+        mandatory_agents: [acceptance-test-engineer]
+        conditional_agents: []
+        parallel_groups: []
+        blocking_agents: []
+        rerun_policy: changed-scope-dependents
+      - id: evidence
+        mandatory_agents: []
+        conditional_agents: [devops-ci-engineer, qa-gatekeeper]
+        parallel_groups: []
+        blocking_agents: [qa-gatekeeper]
+        rerun_policy: changed-scope-dependents
+      - id: review
+        mandatory_agents: [completion-reviewer, qa-gatekeeper]
+        conditional_agents: [implementation-reviewer]
+        parallel_groups: []
+        blocking_agents: [qa-gatekeeper, completion-reviewer]
+        rerun_policy: changed-scope-dependents
+    review_profile: runtime-heavy
+
+  - skill: qfai-prototyping
+    phases:
+      - id: plan
+        mandatory_agents: [delivery-planner]
+        conditional_agents: [product-experience-architect]
+        parallel_groups: []
+        blocking_agents: [delivery-planner]
+        rerun_policy: failed-agents-only
+      - id: build
+        mandatory_agents: []
+        conditional_agents: [frontend-engineer, backend-engineer]
+        parallel_groups:
+          - [frontend-engineer, backend-engineer]
+        blocking_agents: []
+        rerun_policy: changed-scope-dependents
+      - id: evidence
+        mandatory_agents: []
+        conditional_agents: [devops-ci-engineer, qa-gatekeeper]
+        parallel_groups: []
+        blocking_agents: [qa-gatekeeper]
+        rerun_policy: changed-scope-dependents
+      - id: review
+        mandatory_agents: [completion-reviewer]
+        conditional_agents: [product-surface-reviewer]
+        parallel_groups: []
+        blocking_agents: [completion-reviewer]
+        rerun_policy: changed-scope-dependents
+    review_profile: ui-bearing
+
+  - skill: qfai-configure
+    phases:
+      - id: analysis
+        mandatory_agents: [delivery-planner, qa-strategist]
+        conditional_agents: []
+        parallel_groups: []
+        blocking_agents: [delivery-planner]
+        rerun_policy: failed-agents-only
+      - id: config
+        mandatory_agents: [devops-ci-engineer]
+        conditional_agents: []
+        parallel_groups: []
+        blocking_agents: []
+        rerun_policy: changed-scope-dependents
+      - id: review
+        mandatory_agents: [completion-reviewer, qa-gatekeeper]
+        conditional_agents: []
+        parallel_groups: []
+        blocking_agents: [qa-gatekeeper, completion-reviewer]
+        rerun_policy: changed-scope-dependents
+    review_profile: runtime-heavy
+
+  - skill: qfai-verify
+    phases:
+      - id: plan
+        mandatory_agents: [delivery-planner, qa-strategist]
+        conditional_agents: []
+        parallel_groups: []
+        blocking_agents: [delivery-planner]
+        rerun_policy: failed-agents-only
+      - id: execution
+        mandatory_agents: [devops-ci-engineer]
+        conditional_agents: []
+        parallel_groups: []
+        blocking_agents: []
+        rerun_policy: changed-scope-dependents
+      - id: review
+        mandatory_agents: [qa-gatekeeper, completion-reviewer]
+        conditional_agents: [implementation-reviewer]
+        parallel_groups: []
+        blocking_agents: [qa-gatekeeper, completion-reviewer]
+        rerun_policy: changed-scope-dependents
+    review_profile: runtime-heavy

package/assets/init/.qfai/assistant/steering/cli-ux-guidelines.md

@@ -0,0 +1,49 @@
+# CLI UX Guidelines
+
+spec-0013 (CAP-0013) で定義された、`qfai validate` の UI/UX 関連出力ガイドライン。
+
+## Output Format
+
+- Issue メッセージにタイムスタンプやランダム値を含めない（冪等性確保）
+- `## Auto Check Results` — `qfai validate` による自動チェック結果セクション
+- `## Manual Review Results` — product-surface-reviewer による手動レビュー結果セクション
+
+## Error Message Format
+
+```text
+<CODE>: <message> [at <file>[:<line>[:<col>]]]
+```
+
+- `file` は issue に path がある場合のみ付与される
+- `line` / `col` は issue に loc がある場合のみ付与される
+
+例:
+
+- `QFAI-DT-002: Circular reference detected: semantic.color.primary [at .qfai/contracts/design/design-tokens.yaml]`
+- `QFAI-MOCK-002: External URL reference in HTML Mock: https://cdn.example.com/style.css [at .qfai/specs/spec-0001/01_Spec.md]`
+
+> **Note:** `.qfai/contracts/design/design-tokens*.yaml` は **optional supporting artifact** である。init 直後にファイルが存在しなくても異常ではなく、token validator は token file が作成された場合にのみ実行される。
+
+## Severity Decision Matrix
+
+| Category                                              | Error                                                   | Warning                                                                  |
+| ----------------------------------------------------- | ------------------------------------------------------- | ------------------------------------------------------------------------ |
+| Design Token schema violation (if token file present) | `$value` empty, circular ref, YAML parse error          | Unknown `$type`, unknown platform                                        |
+| HTML Mock structure                                   | External URL, script tag, CSS fallback missing          | Missing state variant, contrast ratio below AA, token comment missing    |
+| Mermaid screen flow                                   | —                                                       | v1 migration, unlabeled transition, flowchart declaration, outside fence |
+| BP/AP DB                                              | Invalid ID format, duplicate ID, missing required field | Invalid platform                                                         |
+| Platform detection                                    | —                                                       | Unknown platform, cross-platform detected                                |
+| Consistency                                           | —                                                       | Fallback mismatch, screen alignment missing                              |
+| Research summary                                      | No sources, no apply, missing required sections         | Freshness below threshold                                                |
+| Agent definition                                      | Missing file, missing section, insufficient items       | Missing collaboration statement                                          |
+
+## `--platform` Option
+
+- 許容値: `web`, `windows`, `mobile-ios`, `mobile-android`, `cross-platform`
+- 未指定時: 自動検出（config → project files → fallback to `web`）
+- 未知の値: warning 発行。platform 値は保持しつつ、実質的に common ルールのみ適用
+
+## Known Limitations
+
+- jsdom v26+ は CSS レイアウトをサポートしないため、タッチターゲットサイズチェックはインライン `style` 属性のみ対象
+- クラスベースまたはスタイルシートベースのサイズ指定は、現実装では未検証としてスキップ（追加 info は未出力）

package/assets/init/.qfai/assistant/steering/manifest.md

@@ -2,41 +2,66 @@
 
 ## Product / Mission
 
-- Summary:
-- Value:
-- Evidence:
+- Summary: QFAI is a Quality-First Development Kit (CLI tool) for AI coding agents. It integrates SDD/ATDD/TDD and verifies specs, contracts, and traceability with 50+ validation rules.
+- Value: Detects agent spec drift and hallucinations via objective gates, ensuring quality.
+- Evidence: README.md, packages/qfai/package.json (description), 02_Inception-Deck.md
 
 ## Axioms (Non-negotiable)
 
 - Axioms / principles (non-negotiable):
-- Decision lens (what we optimize for):
-- Evidence:
+  - Correctness > Completeness > Usability > Performance > Extensibility (Trade-off priority)
+  - Validators are pure async functions (no side effects, return Issue[] only)
+  - Layered specs: 1 CAP = 1 spec directory
+  - Drift protocol: Unauthorized editing of upstream artifacts is prohibited
+- Decision lens (what we optimize for): Minimize false positives (reliability), traceability completeness
+- Evidence: 02_Inception-Deck.md (Trade-offs), 09_Constraints.md (TC-09)
 
 ## Compatibility vs Change Rubric
 
-- Criteria (Compatibility):
-- Criteria (Change):
-- Examples:
-- Evidence:
+- Criteria (Compatibility): validate.json is an internal contract (not a stable API). CLI command system follows semver.
+- Criteria (Change): Breaking changes deferred until v2.0. Migration guide required.
+- Examples: `_shared/` -> `_policies/` rename (v1.5.3), spec-pack -> layered migration (v1.4.17)
+- Evidence: CHANGELOG.md, OQ-0003 (validate.json), OQ-0004 (legacy deprecation)
 
 ## Governance (Ownership / Review / Evidence)
 
-- Owner:
-- Review / approval:
-- Evidence requirements:
-- Update cadence:
-- Evidence:
+- Owner: aganesy (maintainer)
+- Review / approval: agent-catalog + agent-routing + review-profiles. Default gate is completion-reviewer; specialist reviewers are conditionally routed by skill/phase.
+- Evidence requirements: evidence file per skill run, validate.log, specs-coverage
+- Update cadence: Per release (semver)
+- Evidence: .qfai/assistant/steering/agent-catalog.yml, .qfai/assistant/steering/agent-routing.yml, .qfai/assistant/steering/review-profiles.yml
 
 ## Evidence
 
-- Rule (e.g., link to repo evidence for every claim):
-- Evidence:
-- Assumptions (if evidence is missing):
+- Rule: All claims must have repo evidence (source code, config, CHANGELOG, discussion-pack)
+  - Evidence:
+    `.qfai/discussion/discussion-20260326072322818/`
+    (v1.7.2 Design Audit & Slop Guardrails),
+    `.qfai/discussion/discussion-20260328120000000/`
+    (v1.7.3 Discussion/UIUX Authoring Foundation),
+    `.qfai/discussion/discussion-20260329120000000/`
+    (v1.7.4 Validation, Review, and Migration Stabilization),
+    `.qfai/discussion/discussion-20260329130000123/`
+    (v1.7.5 Runtime & Evidence Foundation),
+    `.qfai/discussion/discussion-20260329175059391/`
+    (v1.7.6 Critique, Calibration, and Full-Harness Expansion),
+    `.qfai/discussion/discussion-20260329195516830/`
+    (v1.7.6 Audit Remediation — targeted correction pass for 13 issues across P0/P1/P2),
+    `.qfai/discussion/discussion-20260330153902875/`
+    (v1.7.9 Convergence Correction Release — truthful implementation alignment across validate/discussion/prototyping/docs)
+  - Assumptions: None (all verified from repository analysis)
 
 ## Non-goals / Not-now (Optional)
 
--
-- Evidence:
+- IDE plugin / GUI development
+- Plugin architecture (to be reconsidered in v2.0)
+- Automated test generation
+- browser QA full audit / screenshot diff / repair loop / external critique adapter (v1.7.1)
+- auto-fix / rewrite for design findings (v1.7.2)
+- evidence schema versioning detail (deferred to v1.7.6, OQ-0001 of discussion-20260329130000123)
+- browser QA output normalization shape (deferred to v1.7.6, OQ-0002 of discussion-20260329130000123)
+- external critique provider / full-harness orchestration / calibration pack / cost observability / long-running handoff (v1.7.5 out of scope → v1.7.6 IN scope)
+- Evidence: 05_Scope.md (Out of Scope), OQ-0001, OQ-0002, discussion-20260329175059391
 
 ## References (Optional)
 

package/assets/init/.qfai/assistant/steering/product.md

@@ -2,31 +2,73 @@
 
 ## What are we building?
 
-- Summary:
-- Evidence: (e.g., README, docs, issue links)
+- Summary: QFAI - Quality-First Development Kit (CLI) for AI coding agents.
+  Enforces SDD/ATDD/TDD workflows with validation gates via five commands
+  (init, validate, report, doctor, guardrails).
+  Prototyping evidence is produced by the `/qfai-prototyping` skill workflow and consumed by `qfai validate`.
+- Evidence: README.md, packages/qfai/package.json, packages/qfai/src/cli/index.ts
 
 ## Who is the user?
 
 - Personas / roles:
-- Evidence:
+  - AI coding agents (Claude Code, GitHub Copilot, Codex, Anthropic Agents)
+  - QA engineers (quality assurance via validation gates)
+  - Project leads (centralized spec management and traceability)
+  - CI/CD engineers (validation integration into pipelines)
+- Evidence: 02_Inception-Deck.md (Stakeholders)
 
-## What is “success”?
+## What is "success"?
 
 - Success metrics / acceptance definition:
-- Evidence:
+  - All CLI command requirements are defined as REQs
+  - All validation rules (50+) are specified
+  - All traceability edges (US->AC->BR->EX->TC) are defined
+  - Zero errors with qfai validate --fail-on error
+- Evidence: 05_Scope.md (Success Criteria)
 
 ## Non-goals
 
--
-- Evidence:
+- IDE plugin / GUI development
+- Code quality analysis (not a replacement for ESLint/SonarQube)
+- Automated generation of tests themselves
+- Semantic analysis of natural language
+- browser QA full audit / screenshot diff / repair loop / external critique adapter (OUT for v1.7.1)
+- Evidence: 02_Inception-Deck.md (NOT List), 05_Scope.md (Out of Scope)
 
 ## Release posture
 
-- Compatibility policy:
-- Breaking change policy:
-- Evidence:
+- Compatibility policy: semver. Maintain backward compatibility of the CLI command system.
+- Breaking change policy: Breaking changes deferred until v2.0. Migration guide (docs/migrations/) required.
+- Evidence: CHANGELOG.md, 09_Constraints.md (DL-02)
+
+## Milestones
+
+| Version           | Description                                                                                                                                                                                                                                                                                                                                                                    |
+| ----------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
+| v1.5.5            | Spec Diff Protocol (SDP) - Incremental execution support                                                                                                                                                                                                                                                                                                                       |
+| v1.5.6            | Review Agent Enhancement - Devil's Advocate + Pattern Doubler                                                                                                                                                                                                                                                                                                                  |
+| v1.5.7            | UI/UX Definition & Review Framework - Design Token, HTML Mock, Expert Sub-agents                                                                                                                                                                                                                                                                                               |
+| v1.6.0            | Implementation Phase Unification - qfai-implement, test-list.md, Phase 1 Validator                                                                                                                                                                                                                                                                                             |
+| v1.6.1            | Guardrail Hardening - Phase 2 Validator, Report Coverage, Template/Docs Update (完了)                                                                                                                                                                                                                                                                                          |
+| v1.6.2 (完了)     | 開発ツールキット堅牢化 — Sub-agent Roster + Completion/Evidence/Parallel Contracts + Docs/Wrappers/Assets Sync                                                                                                                                                                                                                                                                 |
+| v1.6.3 (完了)     | Copilot レビューインストラクション配布 — qfai init に Copilot レビュー指示テンプレートを統合                                                                                                                                                                                                                                                                                   |
+| v1.6.4 (完了)     | Codex サブエージェント実装 — 39 TOML エージェント + config.toml                                                                                                                                                                                                                                                                                                                |
+| v1.6.5 (完了)     | デザインディレクション＆UI品質強化 — DDP + Navigation/Screen Flow + Render Critique Loop + Fidelity Evaluation                                                                                                                                                                                                                                                                 |
+| v1.7.0 (完了)     | ディスカッション設計強化 — UI-bearing detection + DDS enforcement + competitive reference registry + error-severity validators                                                                                                                                                                                                                                                 |
+| v1.7.1 (完了)     | Render Evidence Automation — render evidence schema (captured/skipped/failed) を定義し、validate/report が structured evidence を消費できるようにする。evidence 生成は skill/workflow 経由                                                                                                                                                                                     |
+| v1.7.2 (完了)     | Design Audit & Slop Guardrails — 静的 design audit 7 dimension + AI slop guardrails SLP-01〜SLP-06 + quality profile severity 制御                                                                                                                                                                                                                                             |
+| v1.7.3 (完了)     | Discussion/UIUX Authoring Foundation — qfai-discussion に uiux/ サイドカーアーティファクト生成、SKILL.md UI-bearing フロー、テンプレート置換・拡張を追加                                                                                                                                                                                                                       |
+| v1.7.4 (SDD 完了) | Validation, Review, and Migration Stabilization — UIX-VAL deterministic validators + UIX-REV semantic reviewers + verify-pack tests + migration support                                                                                                                                                                                                                        |
+| v1.7.5 (完了)     | Runtime & Evidence Foundation — evidence schema recovery + optional render evidence capture + backend provider abstraction + browser QA structured outputs                                                                                                                                                                                                                     |
+| v1.7.6 (完了)     | Critique, Calibration & Full-Harness Expansion — external critique adapter + calibration pack + explicit premium lane (skill-driven) + cost/time observability + handoff artifacts + display/stub detection                                                                                                                                                                    |
+| v1.7.7 (完了)     | Remediation & Prototyping Readiness — static-first prototyping default + full-harness mode exposure + 3-layer eval reconciliation + strategy/contract upgrade + UI-bearing detection fix + render evidence wiring + browser QA findings + doc normalization + migration support                                                                                                |
+| v1.7.8 (完了)     | Canonical Convergence — design taste interview + trend research + 3-layer evaluation convergence + scoring-ready schema + strategy/screen contract upgrade + UI-bearing detection unification + static-first prototyping rewrite + full-harness mode convergence + render evidence wiring + browser QA MVP + reviewer extension + migration normalization + docs normalization |
+| v1.7.9 (完了)     | Convergence Correction Release — canonical validator registration, discussion completion convergence, honest render evidence/browser QA wiring, reviewer routing alignment, docs maturity normalization                                                                                                                                                                        |
+| v1.7.13 (進行中)  | Canonical Sidecar Convergence — selected direction SSOT moved to 31_selected_anchor_screen.md, option comparison remains in 30_option_comparison.md, sidecar-first read order, DDS/anchor vocabulary removal, validator semantics rewrite, template-validator self-consistency                                                                                                 |
 
 ## Open questions
 
-- Blocking:
+- Blocking: none
 - Non-blocking:
+  - OQ-0003: validate.json external API stability (deferred to v2.0)
+  - OQ-0004: Legacy spec-pack deprecation schedule (deferred to v2.0)

package/assets/init/.qfai/assistant/steering/research-first-protocol.md

@@ -0,0 +1,58 @@
+# Research-First Protocol
+
+spec-0013 (CAP-0013) で定義された、UI/UX 専門家サブエージェントの作業冒頭リサーチプロトコル。
+
+## Trigger
+
+`/qfai-discussion` コマンド実行時に自動トリガーされる（DEC-0013-0004）。
+
+## Output Schema
+
+```yaml
+research_summary:
+  sources:
+    - id: string # SRC-XXXX format (必須)
+      title: string # 文献タイトル (必須)
+      url: string # URL (必須)
+      published: string # YYYY-MM-DD (必須)
+  best_practices:
+    - id: string # BP-XXXX format
+      category: string
+      title: string
+      description: string
+      source_id: string # sources[].id への参照
+  anti_patterns:
+    - id: string # AP-XXXX format
+      category: string
+      title: string
+      description: string
+      source_id: string
+  reflection:
+    - source_id: string
+      finding: string
+      action: apply | reject | defer
+      reason: string
+```
+
+## Freshness Rule
+
+- 参照ソースの直近 2 年以内の参照率 ≥80%
+- 80% 未満の場合は freshness warning を発行
+- 古いソースを含める場合は、歴史的重要性などの理由を明記
+
+## Source Citation Rule
+
+- 全エントリに `id`, `title`, `url`, `published` を記録（1 件でも欠落はバリデーションエラー）
+
+## Conflict Protocol
+
+- 新しいリサーチ結果が既存の BP/AP ルールと矛盾する場合:
+  - `reflection[].action: reject` — 新知見を不採用
+  - `reflection[].action: defer` — 判断を延期
+  - **自動上書き禁止**: 既存ルールを自動的に書き換えてはならない
+- `reflection[].action: apply` が 1 件以上必須
+
+## Storage
+
+- `research_summary` はカレントの discussion-pack 内（`## Research Summary` セクション）に記録
+- グローバルには永続保存しない（DEC-0013-0002）